- Lauber, T. Bruce
University of Maine
I conducted a retrospective study to investigate the effects of the U.S. Department of Agriculture's Conservation Reserve Program (CRP) on bird populations. The CRP pays farmers to convert cropland to permanent vegetative cover under long-term contracts and is primarily intended to control crop production and reduce soil erosion. Many species of birds, however, could be affected by the changes in nesting habitat and available food sources as land in agricultural areas is retired.
I conducted my work using data from several computerized data sets. Bird population data were obtained from the U.S. Fish and Wildlife Service's Breeding Bird Survey. Statistics on agricultural land use were obtained from the U.S. Department of Agriculture. Weather and climatic data were derived from statistics kept by the National Climatic Data Center. These data sets provided county-level estimates of density for species of birds and of numerous variables, including the CRP, which could influence the distribution of these species.
Some 102 species associated with agriculture were selected a priori for study. I tested for spatial correlations of these species' distributions with the CRP and other variables using decision tree methodology. This method allowed the detection of non-linear effects and interactions of variables in a straightforward manner. These analyses not only indicated which variables were correlated with species' distributions, but also identified locations where such correlations existed. Some 31 of the 102 species were spatially correlated with the CRP, although all of these correlations could be at least partially attributed to confounding variables.
A consistent geographical pattern of association with the CRP was detected for 12 of the 31 species. For the 10 of 12 species for which sample sizes were sufficient, I tested for population responses after the introduction of the CRP. Four species, Western Meadowlark (P = 0.004), Ring-necked Pheasant (P = 0.084), Brown-headed Cowbird (P = 0.099), and Northern Bobwhite (P = 0.139) showed evidence of positive population responses after the introduction of the CRP.